P6.13
Supercell Differentiation and Organization for the 19 April 1996 Illinois Tornado Outbreak
Bruce D. Lee
Department of Earth Sciences
University of Northern Colorado, Greeley, Colorado
Brian F. Jewett and Robert B. Wilhelmson
Department of Atmospheric Sciences
University of Illinois, Urbana, Illinois
1.
INTRODUCTION
On April 19, 1996, an outbreak of tornadic
supercells struck portions of Illinois and adjacent states
(Storm Data, 4/96). In Illinois alone, 36 tornadoes were
documented, setting a single day record while
additionally exceeding the average number of
tornadoes for an entire year. An estimated $30 million
in damage was reported in 31 counties in Illinois with 1
fatality. Most of the tornadoes on this day were shortlived, but a few persisted for over 20 min and had
damage paths of nearly 25 km with peak intensity rated
at F3. To underscore the scope of this outbreak, Iowa,
Indiana and Missouri reported 2, 13 and 5 tornadoes,
respectively.
A fascinating aspect of this case involves the early
evolution, interaction and merger of cells that initiated
along the dry trough (Martin et al. 1995) just west of the
Mississippi River in Missouri and the warm front that
was located along a northwest-southeast line from
southern Iowa to east central Illinois (Fig. 1). Radar
analysis of this evolving convection presented in section
3 reveals a complex pattern of storm splits and
mergers. Tornadic supercells simultaneously occurred
along both fronts on this day. Of particular interest,
from more than a dozen cells initiating along the dry
trough in northeast Missouri and western Illinois, only 2
large long-lived supercells remained that tracked across
central Illinois, spawning numerous tornadoes. We will
investigate whether a relationship exists between cell
mergers and supercell intensification and the timing of
tornadogenesis. A companion presentation by the
authors (P2.4) will present modeling results that
address the role of the prominent boundaries on this
day in terms of their local environment influence on
storm intensity, structure and organization. It is hoped
that studies such as this will better our understanding of
thunderstorm interactions and local environment effects
as they influence the evolution of supercell
thunderstorms and related phenomena.
_______________________
Corresponding author address: Dr. Bruce D. Lee
Dept. of Earth Sciences (Ross Hall 3320)
Univ. of Northern Colorado, Greeley, CO 80639
email: bdlee@unco.edu
2.
SYNOPTIC ENVIRONMENT
The surface synoptic environment for 2100 UTC
shown in Fig. 1 is dominated by an intensifying surface
cyclone (992 mb) in southwest Iowa that is positioned
under the left exit region of an approaching 65 m s -1
upper-level jet streak. The primary boundaries for
storm initiation on this day were the warm front and dry
trough. A tongue of dew points in excess of 18 oC had
advected northward on strong south-southeast winds in
the morning and early afternoon hours in advance of
the dry trough. In northeast Missouri the dry trough
represents the leading edge of a 16 oC drop in dew
point. Moderate instability existed in the warm sector
with 2300 J kg-1 of CAPE and an LI of -7 oC at Lincoln,
IL based on the 0000 UTC sounding. Strong 3 hr
pressure falls of about 6 mb in western Illinois and
northeast Missouri kept "backed" winds through midafternoon in the warm sector, sustaining strong vertical
shear. The moderate instability, coupled with the strong
low-level vertical shear and triggering presence of the
major boundaries set the stage for a significant severe
storm outbreak. Not shown on Fig. 1 is the presence of
a cold front aloft that may have played a role in deep
convection initiation. This addition factor is being
studied by the authors.
Numerous storms initiated along both of these
boundaries primarily between 2000 and 2100 UTC with
the exception of a small cluster of storms that initiated
about an hour earlier northeast of St. Louis, MO. This
small cluster would gradually amalgamate to become
the largest supercell of the day, travelling isolated down
the warm front through eastern Illinois and across
central Indiana.
The first severe thunderstorm
warnings were issued at about 2100 UTC.
3.
RADAR ANALYSIS
To understand the supercell differentiation and
organization processes active on this day, cells
developing along or near the key boundaries were
tracked and plotted in Fig. 2 using level 2 radar archive
data from Davenport, IA, Lincoln, IL and St. Louis, MO.
To construct this tracking map of peak reflectivity
centroids, cells initiating between 2000 and 2250 UTC
were plotted and tracked until 0100 UTC. Only cells
that reached and maintained at least a reflectivity of 30
L
1.36
10
0
0
H
4.69
L
1.57
1000
8.0
0
H
6.20
H
10.6
0
10
L
992
.
0
L
H
5.13
16
.0
L
-.945
100
0
8
10
0
10
0
.0
16.0
0
-8
.0
0
8.
0
0
.0
H
20.6
100
0
10
0
16.0
8
H
19.5
Fig. 1.
Surface analysis for 2100 UTC on 19 April 1996.
isodrosotherms.
dBZ for a 12 min. period were plotted. In order to focus
on the heart of the outbreak, no cells initiating west of
Des Moines, IA or south of Lebanon, MO after 2210
UTC were plotted. Two prominent flanking line cells
(gray tracks) that initiated at 2350 UTC, which were
deemed very important to the evolution of the 2 central
Illinois supercells, are also included. In all, 81 cells are
tracked.
With respect to the frontal configuration noted in
Fig. 1, it is easy to pick out the convective initiation
zones, mainly along the dry trough and warm front. In
some large outbreaks such as this, a large degree of
complexity is involved as the number of surviving cells,
is reduced to a subset of primarily supercells. Note that
the thunderstorms designated as supercells (bold
tracks) in Fig. 2 demonstrated persistent mesocyclones
and moved to the right of the mean wind shear. The
tracking analysis shows this complexity through an
Solid lines are isobars and dashed lines are
intricate pattern of mergers (M=14), storm splits(S=16)
and other interactions (I=2). Considering the storms
initiating near the dry trough in eastern Missouri, after
the first 2 hours, only a small percentage of the storms
have survived. Many storms initiated in the region of
high moisture convergence near the dry trough, moved
off the trough, and within about 20 - 40 min. simply died
in an environment that apparently was not conducive to
further growth. Other storms are involved in mergers
(Westcott 1984) and the most intense of these early
storms undergo storm splitting (Klemp and Wilhelmson
1978). The splits then interact and, in several cases,
merge with right moving supercells to their immediate
north. By 2300 UTC, of the numerous cells initiating
along the dry trough north of St. Louis, MO, or warm
front east of Ottumwa, IA, 9 supercell storms remain
with few other ordinary (non-rotating) thunderstorms.
Waterloo
Rockford
Cedar Rapids
Des Moines
Davenport
M
I
M
M
Ottumwa
I
S
S
S
S
S
Macomb
Peoria
Bloomington
M
S
S
M
T
M M
S
T
T
Lincoln
T
S
M
T
S
M
M
T
T
Springfield
Champaign
M
Decatur
T
M
M
S
Effingham
Sedalia
Fulton
St. Louis
S
M
Mt. Vernon
S
S
S
S
Farmington
Lebanon
Fig. 2. Cell centroid tracks for the 19 April 1996 outbreak. Bold tracks represent cyclonic supercell storms and
dashed tracks designate storms produced from splitting. S, M, I and T represent a storm split, merger, other
interaction, or tornado report, respectively. Tracks with arrows designate storms that were still underway at 0100
UTC. See text for further details.
It appears differential storm motion was
responsible for some of the merger events whereby a
more mature cyclonically rotating thunderstorm moving
to the right of the mean shear interacts with a younger
cell (Klemp et al. 1980). Another mode of merger, also
based on differential motion, resulted from the
previously mentioned splitting process whereby a left
moving anticyclonically rotating cell intersects a right
moving cyclonically rotating supercell. A third mode of
merger involved significant new convection generated
along the flanking line of a supercell that ultimately
merges with the parent cell.
All three modes of merger may be seen by
following the west-east evolution of the 2 central lllinois
supercells (Bloomington and Springfield cells). The
Bloomington storm experiences 3 merger events while
the Springfield storm experiences 4 merger events.
Thus, numerous cells are gradually consolidated into
just 2 supercells.
Another important product of these cell mergers
involves thunderstorm intensification. The majority of
the merger events for the 2 central Illinois supercells
result in cell intensification as indicated by peak
reflectivity and high reflectivity areal coverage.
Moreover, these supercells generally display enhanced
classic "hook" appendage reflectivity structure during or
just after the cell interaction. An analysis of NEXRAD
wind data to verify the inference from the reflectivity
structure (i.e., that the mesocyclones are strengthened)
is underway. Did the mergers enhance the existing
mesocyclones or did they in fact prompt new
mesocyclone development? Both the Bloomington and
Springfield storms had persistent mesocyclone
presence, but were they the same mesocyclone or new
ones developing concurrent with the merger event?
The first and second authors observed in the field that
the low-level mesocyclone strength on the Springfied
supercell was markedly cyclic. NEXRAD data appears
to confirm this cyclic mesocyclone intensity for both
central Illinois storms.
4.
DISCUSSION
The first phase of this observational study has
described the early convective complexity and cell
interactive processes that produced, in a 2 - 3 hr period
(from initiation), a subset of 9 - 12 supercell
thunderstorms from a large number of initial cells. The
common modes of merger included interactions
resulting from: 1) differential motion created when a
younger or weaker storm approaches a more mature or
stronger cell that has already acquired cyclonic rotation
and is moving to the right of the mean shear, 2)
differential motion created from a left moving
anticyclonically rotating cell intersecting a right moving
cyclonically rotating supercell, and 3) prominent new
cells generated along the parent cell's flanking line that
propagate into the parent supercell.
Cell interaction
resulting from the configuration of line orientation with
respect to the vertical shear vector will also be explored
and compared to the recent idealized simulations of
Bluestein and Weisman (2000).
Further radar analysis will address what the
merging process structurally represents in the 19 April
1996 outbreak. In some cases it appears that the
dominant storm becomes stronger and better
organized, and in other cases, it appears a new updraft
and associated high reflectivity are produced near the
point of merger. Recent model simulations of merging
storms appear consistent with this latter scenario
(Bluestein and Weisman 2000; Finley et al. 2000). On
a related topic, do the mergers lead to enhanced
rotation of a pre-existing mesocyclone or sponsor new
mesocyclogenesis, especially at low-levels?
It
appeared visually (and the tornado reports verified) a
pattern of cyclic intensification of the low-level
mesocyclone.
To see if there was any temporal
relationship between merger events and tornado
occurrence, tornado reports were plotted for the 2
central Illinois supercells up to 0100 UTC. The timing
of tornadogenesis and merger events was striking. For
the Bloomington supercell, 2 of the 3 tornadoes
associated with this storm occurred near the location
and time of a merger. Similarly, for the Springfield
storm, 4 of the 5 tornadoes reported with this storm (up
to 0100 UTC) occurred near a merger event.
The
relationship
between
merger
events
and
tornadogenesis is very compelling.
6.
REFERENCES
Bluestein, H. B., and M. L. Weisman, 2000: The interaction of
numerically simulated supercells initiated along lines. Mon. Wea.
Rev. In press.
Finley, C. A., W. R. Cotton, and R. A. Pielke, 2000: Numerical
simulation of tornadogenesis in a high precipitation supercell. Part I:
Storm evolution and transition into a bow echo. J. Atmos. Sci. In
press.
Klemp, J. B., and R. B. Wilhelmson, 1978: Simulations of right- and leftmoving storms produced through storm splitting. J. Atmos. Sci., 35,
1097-1110.
Klemp, J. B., P. S. Ray, and R. B. Wilhelmson, 1980: Analysis of
merging storms on 20 May 1977. Preprints, 19th Conf. on Radar
Meteorology, 317-324.
Martin, J. E., J. D. Locatelli, P. V. Hobbs, P. Y. Wang, and J. A. Castle,
1995: Structure and evolution of winter cyclones in the central
United States and their effects on the distribution of precipitation.
Part I: A synoptic-scale rainband associated with a dryline and lee
trough. Mon. Wea. Rev., 123, 241-264.
National Climate Data Center, 1997: Storm Data, 38, 4/96 issue.
Westcott, N., 1984: A historical perspective on cloud mergers. Bull.
Amer. Meteor. Soc., 65, 219-226.